Zoho Analytics MCP Server

The Zoho Analytics MCP Server (Beta) implements the Model Context Protocol (MCP), a standardized interface that enables AI models to interact seamlessly with applications. This middleware solution bridges the connection between AI agents and Zoho Analytics, providing powerful data analysis capabilities through a unified interface.

Note: The Zoho Analytics MCP project is currently in its early development (Beta) phase. As we continue to improve and refine its features, the available tools and functionalities may change.

Getting Started

1. Docker Image Setup
2. Configuring Environment Variables
3. Remote MCP Server Setup (Hosted Option)
4. Integrating with MCP Hosts
   • Claude Desktop
   • VSCode
   • Cursor

1. Docker Image Setup

System Requirements To build the Zoho Analytics MCP server from source, you need to have Docker installed on your system. Before setup, please make sure to have docker and a container runtime installed in your device.

Pull Docker Image

If you prefer to use the pre-built Docker image, you can pull it directly from Docker Hub:

docker pull zohoanalytics/mcp-server:latest

2. Configuring Environment Variables

Obtaining OAuth Credentials To configure the Zoho Analytics MCP Server, you need to provide OAuth credentials (Client ID, Client Secret, and Refresh Token) as environment variables. Follow these steps to generate them:

1. Go to the Zoho Developer Console
2. Create a new Self-Client application.
3. Enable the Zoho Analytics API scope.
4. Generate your Refresh Token. For detailed instructions, refer to the official API Authentication Documentation.

Required Environment Variables

Configure these essential variables before integrating with the MCP Hosts:

Other Environment variables

The following is a list of other optional environment variables:

3. Remote MCP Server Setup (Self-Hosted)

Remote MCP lets you host Zoho Analytics MCP once and share it with your organization over HTTP(S), removing local Docker installs, manual OAuth steps, and per-user updates. Authentication, token lifecycle, and upgrades are handled centrally so end users just point their MCP client to a single URL.

Client configuration

Any MCP host that supports the HTTP transport can point to your hosted instance:

{
  "mcpServers": {
    "ZohoAnalyticsMCP": {
      "url": "<hosted-mcp-server-url>",
      "type": "http"
    }
  }
}

or simply add the in the corresponding location the MCP host requires. Check the documentation of your MCP host for more details.

Self-hosting Setup Instructions

1. Register a static OAuth client in the Zoho Developer Console (server-based application). Set the redirect URI to <your-public-url>/auth/callback and record the client ID/secret.
2. Create a .env file for the hosted server with the variables below:

Before configuring and running the Remote MCP server, please check the Security Guidelines for best practices on securing your hosted instance.

1. Run the Remote MCP container on the host machine:

docker run --name remote-mcp-server \
    --network=host \
    --env-file .env \
    zohoanalytics/mcp-server:remote-v1.1

Once the container is running, share MCP_SERVER_PUBLIC_URL with your users; their MCP clients will connect over HTTP without any local setup.

4. Integrate with MCP Hosts

Zoho Analytics MCP Server can be integrated with any MCP host. Below are some sample integrations that demonstrate how this can be done.

Claude Desktop Configuration

To configure the Zoho Analytics MCP server with Claude Desktop, add the following configuration to your Claude settings:

1. Install Claude Desktop in your system.
2. Open the Claude Desktop application and navigate to Settings > Developer tab.
3. Click Edit Config to configure (claude_desktop_config.json) with your MCP server details.

{
  "mcpServers": {
    "ZohoAnalyticsMCP": {
      "command": "docker",
      "args": [
        "run",
        "-e", "ANALYTICS_CLIENT_ID=<YOUR_ANALYTICS_CLIENT_ID>",
        "-e", "ANALYTICS_CLIENT_SECRET=<YOUR_ANALYTICS_CLIENT_SECRET>",
        "-e", "ANALYTICS_REFRESH_TOKEN=<YOUR_ANALYTICS_REFRESH_TOKEN>",
        "-e", "ANALYTICS_ORG_ID=<YOUR_ANALYTICS_ORG_ID>",
        "-e", "ANALYTICS_MCP_DATA_DIR=<YOUR_ANALYTICS_MCP_DATA_DIR>",
        "--network=host",
        "-i",
        "--rm",
        "-v", "<YOUR_ANALYTICS_MCP_DATA_DIR>:<YOUR_ANALYTICS_MCP_DATA_DIR>",
        "zohoanalytics/mcp-server"
      ]
    }
  }
}

VSCode Configuration

To configure the Zoho Analytics MCP server with Visual Studio Code:

1. Install the VSCode in your system.
2. Open the VSCode application and navigate to Settings.
3. In the Settings page, search for MCP configurations. Select the Edit in settings.json option to configure the MCP server.

{
  "servers": {
    "zoho_analytics_mcp_server": {
      "type": "stdio",
      "command": "docker",
      "args": [
        "run",
        "-e", "ANALYTICS_CLIENT_ID=<YOUR_ANALYTICS_CLIENT_ID>",
        "-e", "ANALYTICS_CLIENT_SECRET=<YOUR_ANALYTICS_CLIENT_SECRET>",
        "-e", "ANALYTICS_REFRESH_TOKEN=<YOUR_ANALYTICS_REFRESH_TOKEN>",
        "-e", "ANALYTICS_ORG_ID=<YOUR_ANALYTICS_ORG_ID>",
        "-e", "ANALYTICS_MCP_DATA_DIR=<YOUR_LOCAL_DATA_DIR>",
        "--network=host",
        "-i",
        "--rm",
        "-v", "<YOUR_LOCAL_DATA_DIR>:<YOUR_LOCAL_DATA_DIR>",
        "zohoanalytics/mcp-server"
      ]
    }
  }
}

Cursor Configuration

To configure the Zoho Analytics MCP server with Cursor:

1. Install the Cursor in your system.
2. Open the Cursor application and navigate to Settings > Tools and Integrations.
3. Click the Add Custom MCP menu to enter your server configuration details, and Save.

{
  "mcpServers": {
    "ZohoAnalyticsMCP": {
      "command": "docker",
      "args": [
        "run",
        "-e", "ANALYTICS_CLIENT_ID=<YOUR_ANALYTICS_CLIENT_ID>",
        "-e", "ANALYTICS_CLIENT_SECRET=<YOUR_ANALYTICS_CLIENT_SECRET>",
        "-e", "ANALYTICS_REFRESH_TOKEN=<YOUR_ANALYTICS_REFRESH_TOKEN>",
        "-e", "ANALYTICS_ORG_ID=<YOUR_ANALYTICS_ORG_ID>",
        "-e", "ANALYTICS_MCP_DATA_DIR=<YOUR_ANALYTICS_MCP_DATA_DIR>",
        "--network=host",
        "-i",
        "--rm",
        "-v", "<YOUR_ANALYTICS_MCP_DATA_DIR>:<YOUR_ANALYTICS_MCP_DATA_DIR>",
        "zohoanalytics/mcp-server"
      ]
    }
  }
}

Tools List

The Zoho Analytics MCP server provides various tools for interacting with Zoho Analytics.

Example Prompts

Here are some example prompts illustrating how you can interact with the Zoho Analytics MCP Server and the kinds of responses you might receive:

Example 1: Data Queries

Prompt:

Can you provide me with details of the top 5 best-selling products in the last quarter, including their names, sales figures, and profit margins?

Response:

Sure! Here are the details of the top 5 best-selling products in the last quarter:
1. XPro Smartphone
   - Sales Figures: $500,000
   - Profit Margin: 20%
2. UltraHD 4K Smart TV (55")
   - Sales Figures: $450,000
   - Profit Margin: 25%
3. Wireless Noise-Cancelling Headphones
   - Sales Figures: $419,500
   - Profit Margin: 15%
4. Fit Pro Smartwatch
   - Sales Figures: $350,000
   - Profit Margin: 30%
5. Home Assistant Smart Speaker
   - Sales Figures: $300,700
   - Profit Margin: 18%

Example 2: Entity Creation

Prompt:

Can you create a new table called "Customer Feedback" with the following columns: Customer ID, Product ID, Feedback Date, Rating (1-5), and Comments?

Once done, can you also insert a sample record into the "Customer Feedback" table with the following details: Customer ID: 12345, Product ID: 67890, Feedback Date: 2024-06-01, Rating: 4, Comments: "Great product, but delivery was delayed."

Response:

Table "Customer Feedback" created successfully with the specified columns, and the sample record has been inserted. Please check the table at: analytics.zoho.com/workspaces/<workspace-id>/view/<view-id>

Example 3: Report Creation

Prompt:

Can you create a report that shows the monthly sales for the top 3 products over the last year?

Response:

The report showing the monthly sales for the top 3 products over the last year has been created.

The created report in analytics:

Troubleshooting

For common issues:

1. Authentication Problems:
   • Verify all OAuth credentials
   • Ensure tokens have proper permissions
   • Check token expiration status
2. Connection Issues:
   • Confirm if the Docker container is running
   • Validate network connectivity
   • Check firewall settings
3. Data Center Errors:
   • Verify ANALYTICS_SERVER_URL and ACCOUNTS_SERVER_URL matches your organization's region
   • Ensure the location code is correct (case-sensitive)
4. File System Problems:
   • Confirm if the data directory exists
   • Check directory permissions
   • Verify sufficient disk space